Omnidirectional reference false alarm suppression method based on radar beam forming

Abstract

The invention relates to an omnidirectional reference false alarm suppression method based on radar beamforming, and the method comprises the steps: carrying out the digital beamforming of an AD sampling signal through an FPGA, generating a multi-directional wave beam and an omnidirectional reference wave beam with a weighting coefficient of 1, and enabling the omnidirectional reference wave beam to be used for representing environment fixed noise; and the DSP end performs MTI, MTD and CFAR processing on each beam data, extracts target information exceeding a threshold, compares the target information with the amplitude of the same position of an omnidirectional reference beam, only retains effective targets higher than a reference value, and removes the rest of the target information as false alarms. Through fusion of time domain, frequency domain and spatial domain feature analysis, dynamic noise calibration and multi-dimensional correlation verification are combined, and the false alarm influence of environmental noise is significantly reduced. According to the method, the amplitude characteristic and the multi-dimensional characteristic of the omni-directional wave beam are used for modeling, real-time processing is achieved on a hardware platform, the target detection precision and reliability of the radar in a complex scene are improved, and the method is suitable for high-interference environments such as cities and coastal areas.

Description

Technical Field

[0001] The present invention relates to the technical field of radar false alarm, and in particular to a method for omnidirectional reference suppression false alarm based on radar beamforming. Background Art

[0002] When the current-stage radar works, it is often "surrounded" by complex electromagnetic interference, ground clutter, meteorological clutter, etc. Traditional constant false alarm rate (CFAR) processing means are difficult to cope with various interference patterns, and false alarms occur frequently. For example, in an urban environment, there are reflections from high-rise buildings and radiation from communication base stations; in a coastal scene, there are sea wave echoes and salt fog effects, where interference is confused with target echoes, consuming processing resources and misleading subsequent operations. There is an urgent need to innovate a false alarm suppression strategy method and achieve "seeing clearly through the clouds" relying on the characteristics of the radar itself. Summary of the Invention

[0003] To solve the above technical problems, the present invention provides a method for omnidirectional reference suppression false alarm based on radar beamforming. The method is basically composed of three processing units: multi-beamforming architecture construction, omnidirectional reference signal feature extraction and modeling, and target detection and false alarm discrimination process, including the following steps:

[0004] Step S1: The FPGA performs digital beamforming on the AD sampled signal, generates multi-pointing beams by adjusting the weighting coefficients, and completes the construction of the multi-beamforming architecture;

[0005] Step S2: On the basis of Step S1, the FPGA additionally generates an omnidirectional reference beam with a weighting coefficient of 1, which is used to capture the fixed environmental noise and transmit it to the DSP together with the data of the other beams;

[0006] Step S3: The DSP performs MTI (Moving Target Indication), MTD (Moving Target Detection), and CFAR (Constant False Alarm Rate) processing on the received beam data, extracts the target amplitude and position information exceeding the preset threshold; compares the target amplitude with the amplitude at the same position in the omnidirectional reference beam. If the target amplitude is greater than the amplitude of the omnidirectional reference beam, it is determined as a valid target, otherwise it is excluded as a false alarm;

[0007] Step S4: Loop and execute Steps S1 to S3 until the host computer sends a stop instruction.

[0008] In an embodiment of the present invention, the digital beamforming in Step S1 adopts the Fourier transform beamforming algorithm to generate multiple overlapping wide beams covering a 360° space, and the overlapping area of adjacent beams is not less than 30% of the beam width.

[0009] In one embodiment of the present invention, the weighting coefficient of the omnidirectional reference beam is 1, and its amplitude is used to characterize the average energy level of the environmental fixed noise, and is periodically calibrated through a dynamic update mechanism. The calibration trigger condition is that the variance of the environmental noise energy exceeds a set threshold.

[0010] In one embodiment of the present invention, the CFAR algorithm in step S3 adopts the cell-average constant false alarm rate (CA-CFAR) or the ordered-statistic constant false alarm rate (OS-CFAR), and the threshold value is dynamically adjusted by the radar host computer according to the real-time signal-to-noise ratio.

[0011] In one embodiment of the present invention, in step S3, the comparison between the target amplitude and the omnidirectional reference beam needs to meet the following conditions:

[0012] The magnitude of the target amplitude is consistent with the magnitude of the omnidirectional reference beam amplitude;

[0013] The azimuth error between the target position and the omnidirectional reference beam does not exceed 10% of the beam width.

[0014] In one embodiment of the present invention, the method fuses multi-dimensional features at the DSP side for false alarm discrimination, including time-domain pulse characteristics, frequency-domain spectrum distribution, and spatial-domain arrival angle statistics, and classifies clutter and interference categories through the DBSCAN clustering algorithm.

[0015] In one embodiment of the present invention, the FPGA uses the Xilinx K7 series chip, the DSP uses the TI TMS320C6678 multi-core processor, the AD sampling rate is 2 GHz, and the quantization bit number is 12 bits.

[0016] The above technical solutions of the present invention have the following advantages compared with the prior art: The method for suppressing false alarms with an omnidirectional reference of the present invention applies the theoretical algorithm in actual engineering based on the FPGA and the DSP, and can minimize the influence of environmental fixed noise on target detection during the radar target detection process. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to make the content of the present invention easier to be clearly understood, the present invention will be further described in detail below according to the specific embodiments of the present invention in conjunction with the accompanying drawings.

[0018] Figure 1 It is a design block diagram of the method for suppressing false alarms with an omnidirectional reference for radar beamforming of the present invention.

[0019] Figure 2 It is a data diagram output after the DSP data processing CFAR algorithm of the present invention.

[0020] Figure 3It is a result graph obtained by comparing the data output from the DSP data processing CFAR algorithm of the present invention with a threshold.

[0021] Figure 4 It is a result graph obtained by comparing the data output from the DSP data processing CFAR algorithm of the present invention with a threshold and then with an omnidirectional reference. Specific implementation mode

[0022] As Figure 1 shown, this embodiment provides a method for suppressing false alarms with an omnidirectional reference based on radar beamforming. The method is basically composed of three processing units: multi-beamforming architecture construction, omnidirectional reference signal feature extraction and modeling, and target detection and false alarm discrimination process, and includes the following steps:

[0023] Step S1: The FPGA performs digital beamforming on the AD sampled signal, generates multi-directional beams by adjusting the weighting coefficients, and completes the construction of the multi-beamforming architecture;

[0024] Step S2: On the basis of Step S1, the FPGA additionally generates an omnidirectional reference beam with a weighting coefficient of 1, which is used to capture the ambient fixed noise and transmit it to the DSP together with the data of the other beams;

[0025] Step S3: The DSP performs MTI (Moving Target Indication), MTD (Moving Target Detection), and CFAR (Constant False Alarm Rate) processing on the received beam data, extracts the target amplitude and position information exceeding the preset threshold; compares the target amplitude with the amplitude at the same position in the omnidirectional reference beam. If the target amplitude is greater than the amplitude of the omnidirectional reference beam, it is determined as a valid target, otherwise it is eliminated as a false alarm;

[0026] Step S4: Loop and execute Steps S1 to S3 until the host computer sends a stop instruction.

[0027] Specifically, for the construction of the multi-beamforming architecture: Design the radar to have the ability to switch and form multi-directional beams. In addition to the conventional beam focusing on the target detection area, a set of wide-beam combinations are added for omnidirectional coverage. Using an antenna array (uniform / non-uniform arrangement), by adjusting the weighting coefficients, control the beam direction, width, and gain to achieve spatial sampling within the radar scanning range. For example, in a phased array, according to the Fourier transform beamforming algorithm, generate narrow beams with a fixed beam width, ensure that the beams cover all angles without signal "dead zones", synchronously collect and time-share store the data of each beam, and construct an omnidirectional raw signal set.

[0028] Furthermore, the Fourier transform beamforming realizes signal focusing through array weighting, and the formula is

[0029]

[0030] Where: M: Total number of array elements

[0031] ω n : The normalized weight of the nth array element, usually

[0032] S n (t): The time-domain signal received by the nth array element

[0033] k = 2π / λ: Wavenumber (λ is the wavelength)

[0034] d: Array element spacing. For phased arrays, d = λ / 2 is selected

[0035] θ: Beam pointing angle. θ ≈ λ(2Md)

[0036] In a specific working example: The radar has a total of 48 beams, each beam has a beam width of 2°, the horizontal direction angle is 90°, and the last beam stores omnidirectional reference data.

[0037] Specifically, for the extraction and modeling of omnidirectional reference signals: After beamforming is completed, the weight coefficients of the formed beams are adjusted to form the extraction and modeling of omnidirectional reference features. The omnidirectional reference needs to be able to simulate the actual environmental noise to provide assistance for subsequent processing. Usually, it is simply to set the beam weighting coefficient to 1 as the default environmental noise reference, or clutter maps are also used to simulate the actual environmental noise to complete the modeling of the reference signal.

[0038] Specifically, the target detection and false alarm discrimination process: The echoes of the conventional detection beams are processed in segments according to a fixed duration, and the features are extracted and compared with the omnidirectional reference model. In the time domain, signals that do not conform to the target pulse law are marked; in the frequency domain, those that overlap with the interference model frequency band and have energy matching are suspicious. Through multi-dimensional fusion determination, suspected false alarms are down-weighted or filtered out. The remaining signals are confirmed as real targets through multi-frame association (speed and azimuth coherence), and then associated tracking output is performed.

[0039] At the same time, as Figure 2 shown, the DSP processes the beamforming data uploaded by the FPGA using a one-dimensional CFAR algorithm, and then outputs the results. Among them, the abscissa is the distance (m*10), and the ordinate is the amplitude (dB). After CFAR algorithm processing, it can be seen that there is still a lot of noise. Next, a reasonable target threshold needs to be set to detect the target.

[0040] As Figure 3As shown, where the abscissa is the distance (m*10) and the ordinate is the number of times exceeding the threshold (times). After the CFAR algorithm is processed and compared with the radar threshold sent by the host computer, those exceeding the target threshold are regarded as valid targets. The number of comparisons is set to 48 times. Each time the threshold is exceeded, a count of 1 is made at that position, and a count of 0 is made at positions not exceeding the threshold. According to the number of times exceeding the threshold, it can be determined where there are targets. At this step, a part of the clutter can be filtered out, but the number of reported targets is still relatively large.

[0041] As Figure 4 shown, for the targets obtained by comparing the thresholds, the amplitudes of the current targets are compared again with the data at the same positions in the omnidirectional reference, and the target data that cannot exceed the omnidirectional reference is eliminated, which can further remove the false alarms formed by the environmental fixed noise. Where the abscissa is the distance (m*10) and the ordinate is the number of times exceeding the omnidirectional reference (times), the accurate target positions are more prominent and the false alarms are further controlled.

[0042] In summary, the result of comparing the omnidirectional reference with the CFAR target result can effectively suppress false alarms.

[0043] Furthermore, for the comparison of the effects with and without the omnidirectional reference, two-level arrays are set as external variables in the software. In the CCS compilation environment, monitoring can be added for viewing. See the following table. cfar_result is the result of comparing with the threshold after the CFAR algorithm, and target_result is the result after adding the omnidirectional reference comparison to the cfar_result data:

[0044]

[0045]

[0046] It can be seen from the table that after comparing the cfar algorithm with the threshold, 30 targets are detected. After comparing with the omnidirectional reference, after removing the environmental noise, there are still 3 reported targets, and the real targets are two. It can be seen that after the omnidirectional reference participates in the target detection, most of the environmental noise is effectively filtered out.

[0047] To verify the feasibility of the method in this embodiment, this method is applied on the hardware platform. From Figure 3 、 Figure 4 it can be seen that using the omnidirectional reference to suppress false alarms has a significant effect in actual engineering applications.

[0048] Obviously, the above embodiments are only examples clearly described and not limitations on the implementation manners. For those of ordinary skill in the art, other different forms of changes or modifications can be made based on the above description. It is not necessary and impossible to list all the implementation manners here. And the obvious changes or modifications derived therefrom are still within the protection scope of the present invention.

Claims

1. A method for omnidirectional reference suppression and false alarm reduction based on radar beamforming, characterized in that, It includes the following steps: Step S1: The FPGA performs digital beamforming on the AD sampling signal, generates multi-pointing beams by adjusting the weighting coefficients, and completes the construction of the multi-beamforming architecture; Step S2: Based on Step S1, the FPGA additionally generates an omnidirectional reference beam with a weighting coefficient of 1, which is used to capture the ambient fixed noise and transmit it to the DSP together with the data of the other beams; Step S3: The DSP performs MTI, MTD, and CFAR processing on the received beam data, and extracts the target amplitude and position information exceeding the preset threshold; Compare the target amplitude with the amplitude at the same position in the omnidirectional reference beam. If the target amplitude is greater than the amplitude of the omnidirectional reference beam, it is determined as a valid target, otherwise it is excluded as a false alarm; Step S4: Loop and execute Steps S1 to S3 until the host computer sends a stop instruction.

2. The method for suppressing false alarms of omnidirectional reference according to claim 1, characterized in that: The digital beamforming in Step S1 adopts the Fourier transform beamforming algorithm to generate multiple overlapping wide beams covering a 360° space, and the overlapping area of adjacent beams is not less than 30% of the beam width.

3. The method for suppressing false alarms of omnidirectional reference according to claim 1, characterized in that: The weighting coefficient of the omnidirectional reference beam is 1, and its amplitude is used to characterize the average energy level of the ambient fixed noise, and is periodically calibrated through a dynamic update mechanism. The calibration trigger condition is that the environmental noise energy variance exceeds the set threshold.

4. The method for suppressing false alarms of omnidirectional reference according to claim 1, characterized in that: The CFAR algorithm in Step S3 adopts the cell-averaging constant false alarm rate or the ordered statistics constant false alarm rate, and the threshold value is dynamically adjusted by the radar host computer according to the real-time signal-to-noise ratio.

5. The method for suppressing false alarms of omnidirectional reference according to claim 1, wherein: In Step S3, the comparison between the target amplitude and the omnidirectional reference beam needs to meet the following conditions: The magnitude of the target amplitude is consistent with the magnitude of the amplitude of the omnidirectional reference beam; The azimuth error between the target position and the omnidirectional reference beam does not exceed 10% of the beam width.

6. The method for suppressing false alarms of omnidirectional reference according to claim 1, characterized in that: The method fuses multi-dimensional features at the DSP end for false alarm discrimination, including time-domain pulse characteristics, frequency-domain spectrum distribution, and spatial domain arrival angle statistics, and classifies clutter and interference categories through the DBSCAN clustering algorithm.

7. The method for suppressing false alarms of omnidirectional reference according to claim 1, wherein: The FPGA adopts the Xilinx K7 series chip, the DSP adopts the TI TMS320C6678 multi-core processor, the AD sampling rate is 2 GHz, and the quantization bit number is 12 bits.

Images